It is common practice in the fiber glass industry today to control the bushings in which molten glass is contained and through which glass fiber formation occurs and to control the bushing which is essentially a heating element by utilizing electrical control devices. Thus, in U.S. Pat. Nos. 4,546,485 and 4,594,087, two systems are described which generally speaking conform to systems currently in use today for producing fiber glass from fiber glass bushings.
In the formation of glass fibers from a bushing, the bushing goes through an operational cycle which involves starting up the bushing, running the bushing, doffing the product wound from the bushing and restarting the bushing. What this means in real terms is that the bushing is subjected to many changes in its thermal history over each running cycle.
A bushing for producing glass fibers is typically constructed of non-reactive refractory metal such as platinum, platinum-rhodium alloy being the preferred metal. The bottom of the bushing is typically divided into a plurality of rows of orifices through which molten glass can readily flow. The orifices usually have on the bottom side of them an associated orifice tip in communication with the orifice so that the molten glass passing through the orifices flows through the tips. Fibers are formed as the molten glass flowing from the tips to the atmosphere is cooled. Fibers formed from the bushing are typically gathered into one or more strands and are attenuated by connecting the strand or strands to the surface of a rotating winder which rotates at sufficient revolutions per minute to pull the strands at linear speeds of 3,000 to 20,000 feet per minute or more.
In operating a glass fiber forming bushing, therefore, molten glass is permitted to flow by Poiseuille's Law through the orifices in the bottom of the bushing. The resulting streams of molten glass are cooled to form filaments as they leave the bushing bottom. Cooling is accomplished by water sprays and environmental air. The fibers or filaments are gathered into one or more strands, usually by placing the filaments as they emerge from the bushing tips into a grooved graphite gathering shoe. The resulting strand or strands are then wound around the surface of the winder and the winder is rotated. Thus the strands are drawn from the bushing by being wound on the winder as it begins to rotate. The winder increases in speed until it reaches the desired speed that will produce a filament of a given diameter based on the diameter of the orifices in the bushing through which the glass is drawn. Another parameter that controls the diameter of the filaments as they leave the orifices is the viscosity of the glass and that is determined by the temperature of the bushing and the glass composition. Since molten glass is continuously maintained in and passed through the bushing during fiber formation, the bushing is fed molten glass through an opening in a forehearth connected to a glass melting furnace.
In the starting up and stopping operation of a bushing, many transient effects occur. Thus, during start up, the initial strands wound on the winder surface are being wound at an accelerating speed which starts out from zero and gradually works its self up to the rotational speed necessary to produce a given filament size. This running of strand at high speed draws environmental air into the filament forming zone and then downwardly at considerable velocity. A bushing then runs for a significant period of time, usually 10 to 30 minutes or longer and the filaments being formed are wound in strand form on the surface of the winder at the desired filament diameter. When the desired weight of material has been accumulated on the surface of the winder the winder is then shut down. This shut down involves a deceleration of the rotation of the winder and a reduction in strand speed. Air flows around the bushing change rapidly as a result and the loss of cooling by the high velocity air present during running results in increased bushing temperatures if all things remain the same except the winder shut down.
As has been previously pointed out, the bushings are controlled by a temperature controller which feeds a signal corresponding to the desired set point for that bushing to the power pack that supplies the bushing current. Thus, for a given viscosity of glass desired from a bushing, it might require for example, a bushing temperature of 2200.degree. F. In such an instance, it is desirable for that bushing to be forced to operate at 2200.degree. F. so that the proper glass viscosity is maintained by the bushing. This, coupled with a control of the rotational speed of the winder through its motor controls accurately and efficiently the filament diameter.
To insure that the controller is operating a bushing at its desired temperature, thermocouples are placed in the sides of the bushing near the bottom. The readings from the thermocouples are then averaged and the resulting signal is sent to the power pack controller feeding current to the bushing. The thermocouples usually are located slightly inboard of the ends of the bushing and close to the bottom on the front wall, i.e., the wall closest to the operator. The thermocouple measurements taken are then passed through a temperature averaging device such as shown in U.S. Pat. No. 4,546,485 to determine the bushing bottom or faceplate temperature. As used herein, the terms faceplate, tip plate and bottom are synonymous. The signal resulting from this average temperature determination is then passed to the controller and the controller forces the bushing to adjust itself to the set point temperature based on the reading it obtains.
It has been found that while bushing controllers can to some degree control the bushing with a certain amount of accuracy, several serious defects are prevalent in this system. First, by taking bushing measurements from the sides of the bushings, close to the bottom at two locations, the effects of noise can be as high as 12% of the temperature signal read, therefore, the signal is inaccurate at least to that degree, i.e., only 88 percent of the signal represents true temperature. The other 12 percent measurad is caused by changing environmental effects and inappropriate thermocouple placement. It has also been found that during the start up, running and doffing of the forming packages that the temperature of the bushings varies over a very wide range and quite rapidly. Despite the accuracy of the thermocouples measuring the temperature of the tip plate, the signals that are generated thereby and fed to the controller contain false signals, i.e., noise. This is especially true when consideration is given to the strand which is an aggregate of the filaments. Thus, to control the yardage of the strand, the temperature of the entire array of tips must, therefore, be controlled. Thus, the effects of things such as environmental air changes occurring near the tip plate, movement of the strands from the gathering shoes to pull rolls during doffing, and other similar occurrences cause rapid temperature changes and gradients which give rise to tip plate measurements that are not a true indication of the thermal condition of the tip plate. While the controller tries to keep the temperature of the bushing constant based on the thermocouple readings, it has been found that these signals are not always representative of the tip plate temperature and therefore, a need for more accurate determination exists.
Still further, situations can arise where a temperature measuring connection to the bushing can become disengaged. When this occurs, the dynamic range of the controller's amplifier is exceeded and a runaway condition can result unless automatic limits on the control system are in place. A runaway could result in temperatures exceeding the softening temperature of the metal used in the leaching in such an instance. In another situation adjustments to fin coolers used to cool the glass streams issuing from bushing tips by an operation can cause a short which, depending on the location of the grounding, can cause common mode signals which, if not impeded, can cause the bushing controller to automatically adjust excessive current flows to the bushing.
The instant invention involves methods and apparatus which substantially reduces the effects of these problems or eliminates them altogether.